A Nonparametric Approach for Detecting Differential Alternative Splicing in RNA-Seq Data

• Activity Number: 409
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 2:45 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Yang Shi*+ and Hui Jiang
• Companies: University of Michigan and University of Michigan
• Keywords: non-parametric ; permutation ; RNA-seq ; alternative splicing ; differential expression ; gene expression

Abstract

High-throughput sequencing of transcriptomes (RNA-Seq) has rapidly evolved as a powerful tool for the study of gene expression and alternative splicing in humans and model organisms. With the reduction of the cost of sequencing, researchers are able to design complicated RNA-Seq experiments and generate large-scale RNA-seq data with hundreds or thousands of samples. We present a non-parametric approach to detect differential splicing of alternative isoforms for large sample RNA-Seq data. Our approach is more robust to outliers, which often exist in large sample size RNA-seq experiments. Simulations studies show our approach has well-controlled type-I error rate and good power in detecting differential splicing events. We also compare our approach with other parametric methods in simulations and a real RNA-seq dataset from prostate cancer patients.